Background: FOLFIRI + bevacizumab (FOLFIRI + BEV) is a standard second-line chemotherapy (Cx) for metastatic colorectal cancer (mCRC) patients (pts) who are refractory or intolerant to fluoropyrimidines (FPs) and oxaliplatin (OX). However, the efficacy of continuing FPs as second-line Cx for pts who are refractory to FPs in first-line Cx remains unclear. Methods: We retrospectively evaluated mCRC pts who received irinotecan (IRI) + BEV or FOLFIRI + BEV as second-line Cx at a single institution from Jan 2010 to Apr 2020. The main eligibility criteria were ECOG performance status (PS) of 0-2, known KRAS status, a standard initial dose of fluorouracil (5-FU) (bolus 400 mg/m², infusional 2400 mg/m²) in FOLFIRI + BEV and IRI (150 mg/m²), refractory to FPs, no prior use of IRI, and prior use of OX. We compared the efficacy and safety of IRI + BEV with those of FOLFIRI + BEV. The adjusted hazard ratios (aHRs) for progression-free survival (PFS) and overall survival (OS) were calculated using a multivariate Cox model that contained variables with p < 0.05 in the univariate analysis to reduce the imbalance between both the treatments. Results: Among the 261 pts, 107 were eligible (IRI + BEV/FOLFIRI + BEV, 31/76 pts). Pt characteristics were as follows (IRI + BEV/FOLFIRI + BEV): median age, 67/62; ECOG PS, 1–2, 55/46%; KRAS mutant, 45/50%; BRAF V600E mutant, 6/11%; right-sided tumor, 19/43%; lactate dehydrogenase (LDH) ≥ 400 U/L, 13/12%; PFS of first-line Cx < 6 month (m), 39/30%; prior use of BEV, 84/63%. Relative dose intensity (RDI) (IRI + BEV/FOLFIRI + BEV) of IRI and BEV was similar in the groups; median RDI (range) of IRI, 80 (44–100) /83 (49–100) %; p = 0.560; median RDI of BEV, 86 (35–100) /83 (20–100) %; p = 0.681. Efficacies (IRI + BEV/FOLFIRI + BEV) after a median follow-up of 13.1/14.3 m were as follows: median PFS, 6.4/5.8 m (HR, 0.90; 95% CI, 0.57–1.38; p = 0.64; aHR, 0.82; 95% CI, 0.50-1.34; p = 0.44); median OS, 16.6/16.5 m (HR, 0.83; 95% CI, 0.51-1.32; p = 0.44; aHR, 1.01; 95% CI, 0.59-1.69; p = 0.97); objective response rate, 25.9/11.3%. Adjustment factors for PFS were prior colorectomy, number of metastatic sites, ECOG PS, liver metastasis, and the level of LDH, while those for OS were prior colorectomy, number of metastatic sites, liver metastasis, peritoneal metastasis, and the level of LDH. All subgroup analyses for PFS and OS according to the pt characteristics also showed no significant differences in the groups. All grade nausea (32/58%) and stomatitis (13/36%) and grade 3–4 neutropenia (23/58%) and febrile neutropenia (0/3%) were less common in the IRI + BEV than in the FOLFIRI + BEV group. Conclusions: Our study suggests that omitting 5-FU from FOLFIRI + BEV as second-line Cx for mCRC pts who are refractory to FPs may lower the occurrence of adverse events without impairing the treatment efficacy.